Haonan Liu, Xiaoyu Li,Li Jin, Wen Shi, Yang Bai, Linhao Zhang, and Hongqi Wang
Cross-layer graph contrastive learning, multi-UAV collaborativenetworks, graph neural networks, dynamic community detection∗ Aerospace Information Research Institute, ChineseAcademy of Sciences, Beijing 100190, China; e-mail: [email protected]; [email protected]∗∗ School of Electronic, Electrical and Communication Engineer-ing, University of Chinese Academy of Sciences, Beijing 100190,China∗∗∗ Key Laboratory of Target Cognition and ApplicationTechnolog
UAVs (unmanned aerial vehicle) are widely employed in disaster relief, agricultural monitoring, logistics distribution, and military reconnaissance. With the expansion of UAV network scales, there are heightened demands for task allocation, path optimisation, and system robustness. The intrinsic community structures within UAV networks play a vital role in enhancing task efficiency and network performance. In order to use the potential community of UAVs to assist in scheduling UAV network scheduling, we propose a novel semi-supervised community construction method based on iterative cross-layer graph contrastive learning. This method integrates graph-level and node-level information to identify key nodes and community structures, thereby optimising task allocation, path planning, and system performance. Experimental results on various datasets demonstrate the methods efficiency in handling large-scale UAV network data and meeting real-time and robustness requirements. It holds significant potential in the realms of intelligent UAV collaboration and automated control.
Important Links:
Go Back
